The Adolescent Component of the National Comorbidity Study (NCS-A) is a combined probability and school-based multi-ethnic sample of 10,000 US adolescents. Dr. David Offord and his team at McMaster University and our research group at Yale adapted the adult diagnostic interview for adolescents. This study provides the first national data on the prevalence and correlates of mental disorder in youth in the general population. Our research team has worked collaboratively with other national experts to select a series of measures of risk factors for major mental disorders. Our research group devoted substantial effort to the design and implementation of the study and the analyses. First, our staff at the NIMH supervised the clinical validation of the fully structured Composite International Diagnostic Interview compared to clinical diagnostic interviews with the semi-structured clinical diagnostic interview for adolescents and children, the K-SADS. Several of our clinical staff and six clinical interviewers also conducted the validity study of 100 adolescents with bipolar disorder or conduct disorder using both the K-SADS and the Family Study Interview for Physical and Mental Health developed by our research group. As part of this study, we developed methods for conducting clinical appraisals of highly structured diagnostic interviews administered by lay interviewers that appear to be an optimal method for conducting reliable and valid interviews of children and adolescents in the general population. We also reviewed more than 200 audiotapes and diagnostic interviews of parents and youth to provide independent ratings of the interviews. Data analysts in our group also spent a substantial effort to error checking, developing diagnostic algorithms, analyzing the validity study, and preparing other measures for data analysis. In order to test our ability to collect biologic samples in large population-based studies, we also developed a study to collect saliva from the entire sample both before and after the interview to assess cortisol, DHEA, and testosterone levels. Preliminary analyses of the first 2000 cases show that baseline cortisol levels are equal for males and females, increase with age and exhibit large diurnal changes. We also found that children with ADHD have lower baseline cortisol and that youth with conduct problems or substance use disorders have less of a decrease in cortisol across the duration of the interview than those without these conditions. We are currently working with experts at the National Institute for Child Health and Development (Dr. George Chrousos) and at the University of Oregon (Dr. Judy Cameron) to determine assays and analyses for the remaining subjects in this study.